diyAudio (
-   Solid State (
-   -   Compensation and Subjectivism.. (

vynuhl.addict 21st April 2007 05:35 AM

Compensation and Subjectivism..
I am happy to be back again browsing the threads. While being quite busy with family, work and music I have found the time over the past few weeks to discover some new sbujective insights, atleast to myself that is. As most of you reading know I was determined to build a blameless amplifier ala self that squashed all bad thoughts on this amplifier and its sonic merits and i thought i did it 100% with the previous feedback/compensation mods but being the way i am and never able to sit still for to long it wasnt the best it could be. I had been using 10pf lag comp and 33pf from vas output to negative input, while this gave a really nice sonic character i found over time the bass while very much there was sort of detached from it all and a little loose, flabby so to speak. I have since pinned this down to the high lead compensation value of33pf. I raised the lag compensation to22pf and used 10pf lead instead and this seemed to bring back the attack to notes that was being smoothed over and nice sounding. I have to say I have never modified and made changes that affect sound quality so profoundly as making changes withing the vas stage architecture. qith 10pf lag lowering the lead value gave a very slight hardly noticeable buzzing on strong solo vocals. I have a torture track that i used for all of these subjective changes and its a accapella hardly compressed solo vaocal track by Alanis Morissette, now I cant say that i am her biggest fan but when she belts out she has a voice that will test any existing honking annoyance factor that may exist deep within an audio system. Needless to say so far she sounds in good health with no clothespins on her nose.


AndrewT 21st April 2007 05:15 PM

I wonder if the phase margin is a little low, giving rise to a touch of overshoot on transient pulses?
This will sharpen up treble but not show in elevated distortion measurements.

JLH used a similar method and he described a setup procedure that used a second RC compensation route parallel to the first with a pot that he suggested be trimmed to square off the test signal and then replaced with a fixed value.

AKSA 22nd April 2007 03:38 AM


Concur 100%!


destroyer X 22nd April 2007 03:45 AM

Perfect Vinyl!...just perfect!
You got it! found the one! you have one of the 'secrets of our universe".

Very good man!


vynuhl.addict 22nd April 2007 05:25 AM

Thanks Hugh, Carlos, I am consistently chasing that being there sonics and am very happy at this point, its very there. Funny enough my better half came home last night and asked me what I did due to the fact it just sounded really good to her, other than that she could care less about it, lol.. Its 10:30 pm here and i havent been able to stop listening to cds since 8pm..Carlos I used your interstation fm trick, your a valuable wealth of hands on info..



AKSA 22nd April 2007 06:16 AM


The interesting thing is that having pursued this subjectively, by ear so to speak, you have discovered something which would have evaded you if you had studied PSpice and striven to reduce distortion. I think this happens because we are listening to music, not test tones.

This is the underestimated area amp design which no one to my knowledge has satisfactorily explained. It's not magic, as some would have it, it's just not completely understood yet.

Now you have found one weak spot, there are more to be found in the input stage. But they are pretty tricky to find; try googling Gerard Perrot.



Dave 22nd April 2007 06:33 AM

This quote is from

"Memory Distortion was discovered a long time ago by a French man called Gérard Perrot, who has since created his own company called Lavardin, which makes amplifiers and other Audio devices."

Nordic 22nd April 2007 07:48 AM

Love the link Dave.

destroyer X 22nd April 2007 09:58 AM

The variable capacitor to tune things were used and suggested by Graham Maynard.

I am the one, that is honored with his attention.



GK 22nd April 2007 10:19 AM

Having nothing better to do this evening, I will attempt, although probably pointlessly, to inject some sanity into this thread:

Frequency compensation is absolutely critical to an amplifiers performance. You simply cannot make willy-nilly changes to critical frequency compensation components and properly evaluate the complete impact of such changes purely by ear. Amongst other things, your ears and an interpretation of what you hear alone cannot be relied upon to tell you with certainty if your amplifier is continuously oscillating supersonically or not, or if the transient response is marginal or if the global loop compensation is not conditionally stable.
Things such as these can drastically alter the sonic performance of the amplifier either directly or indirectly, and without the appropriate measuring equipment required and a solid theoretical knowledge of how to empirically optimise an amplifiers frequency compensation, you are simply poking around in the dark.

You have reported a drastic change in the sonic performance of your amplifier with which seems, by your description, to be a rather unscientific approach to changing compensation component values.
By virtue of the latter, the possibility of the former (a drastic variation in audio quality) is not even remotely impossible, mysterious or surprising.
The only thing that your subjective experience (assuming that you have reported it here honestly and are not under delusion) suggests is that there is something critically wrong with your amplifiers frequency compensation in one form of alteration or the other.
And nobody here can say with any authority whatsoever exactly what it may be without, at the very least, thoroughly analysing your circuit - and least of all declare that you have stumbled upon a rather magical “secret-of-the(your?)-universe” ratio of lead to lag compensation capacitor values that will miraculously transform the sonic performance of an otherwise adequately compensated and properly designed amplifier.


All times are GMT. The time now is 08:48 AM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright ©1999-2017 diyAudio